| In the first chapter of this dissertation, I estimate the long-run price elasticity of supply of renewable electricity generating capacity using the variation in state mandates for renewable electricity production. These mandates require electricity providers to purchase a predetermined fraction of their electricity from renewable sources and typically increase annually in each state. Using the variation that these mandates induce in generating capacity, I use an instrumental variables strategy to estimate the price elasticity of supply. My estimate of 2.7 translates into a cost of carbon dioxide abatement of at least ;The second chapter, written jointly with Paul Courant, David Mendez, and Kenneth Warner, conducts a benefit-cost analysis on the Environmental Protection Agency's guidelines for residential radon remediation. We use an agent based approach that adds important dimensions of heterogeneity to the analysis. This allows us to estimate the likely capitalization of the capital cost of remediation technology into housing prices. We find that most households are better off by not paying the annual cost of remediation and that only the least mobile households with smokers in high radon concentration houses would undertake the capital cost of remediation. Since only a small fraction of the population values radon remediation, our model suggests that approximately 10% of the capital cost is capitalized into the resale value of the house.;The third chapter examines the finite sample and distributional properties of the nested fixed point algorithm. Starting from the basic setup described in John Rust's 1987 paper, I simulate data sets with varying sizes and distributional assumptions on the unobserved component of the model. I find that even in sample sizes of up to 8,000 observations, the nested fixed point algorithm can display finite sample bias and variances substantially larger than the theoretical asymptotic variance. This is also true with departures from distributional assumptions, with the mean square error increasing by a factor of 10 for some distributions of unobserved variables. |
